A Comprehensive FISH- and Array-Based Diagnostic Screeningstrategy for the Assessment of Clinically Relevant Genetic Parameters: A Prospective Analysis of 205 Childhood ALL Cases

Background: The detailed definition of causative genomic alterations is not only an indispensable prerequisite for the predictive and prognostic subdivision of childhood acute lymphoblastic leukemia (ALL) but increasingly also one, on which individualized treatment approaches will be based on. Apart from the already well-established genetic categories, the recent identification of several new classes of potentially relevant alterations together with the increasing availability of novel therapeutic options therefore necessitates a diagnostic workflow that is able to satisfy the ensuing clinical needs in a comprehensive manner. The two most interesting changes in this context are the therapeutically targetable recurrent but rare and heterogeneous tyrosine kinase and JAK2 pathway-activating (TKA) gene fusions and the more elusive cohort of apparently relapse risk-prone cases with hitherto only vaguely defined combinations of gene region-specific copy number alterations (CNA). Despite the availability of a multitude of applicable techniques, the fast and cost-efficient identification of the entire expectable abnormality patterns still remains a challenge, especially if one needs to perform the diagnostic work-up on a single case basis. We previously proposed that these diagnostic requirements could be covered best with a systematic hierarchical FISH screening approach for the identification of gene fusions together with array (combined SNP and non-polymorphic probes) analyses of genome-wide quantitative and qualitative large- and small-scale copy number aberrations (CNAs).

Material and Methods: Since June 2015 we have therefore evaluated the feasibility of such a workflow in a prospective manner and screened so far 205 patients (i.e. 184 with B- and 21 with T-ALL, including 18 relapses with 14 diagnostic/relapse pairs) that were consecutively enrolled in the Austrian AIEOP-BFM 2009 treatment study. Cytogenetic preparations served as backup, since metaphase spreads were used for further FISH clarification of otherwise unresolvable complex rearrangements or ploidy patterns if deemed necessary. All identified gene fusions were subsequently validated with single or multiplex RT-PCR analyses.

Results: FISH screening was positive in 90% (184/205) of cases and provided already the most essential diagnostic clues. CNAs were present in all T-ALL and 97% (179/184) of B-ALL cases, including 13 with an IKZF1^pluspattern and three with ERG deletions, which both will be used das stratifying markers in the upcoming treatment trial. Taken together, our screening strategy allowed the unambiguous classification of the vast majority of B-ALLs: 70 hyperdiploid, 3 hypodiploid, 35 ETV6-RUNX1, 6 KMT2A-rearranged, 8 TCF3-PBX1, 2 BCR-ABL1, 4 dic(9;20), 5 iAMP21, 8 IGH-rearranged, 6 P2RY8-CRLF2, 3 PAX5-rearranged, 3 ZNF384-rearranged, 2 ETV6-rearranged, 3 TKA fusion-positive and 10 so called "B-other" cases without any apparent specific abnormality. RNA-sequencing analyses of these ten cases revealed that seven of them harbored a DUX4 gene rearrangement.

Conclusions: Apart from its proven practical diagnostic value, our combined FISH/array approach has also several additional advantages, especially if one considers the amount of achievable information. Both procedures require only little amount of material and are highly standardized, reproducible and robust technologies. Moreover, arrays deliver DNA-sequence-based data, the coordinates of which can be efficiently stored, processed and analyzed. As such, they not only serve as a pure diagnostic tool but also as a valuable discovery platform.